In this report, BERA’s Expert Panel on Assessment argues that school league tables and SATs should be abolished in favour of a new, fairer and more sustainable accountability system. It sets out proposals to replace high-stakes tests and school league tables with a nationally representative sample of pupils who would be tested over time on a broader range of skills and competences. It would also involve gathering a new range of information on schools and their communities using school, parent and pupil questionnaires. This richer combination of data could then be used to inform investment and other aspects of system improvement.

This new system would be developed with the involvement of the profession, parents and pupils, be overseen by a new independent body, and provide clearer insights into educational inequalities which have been thrown into sharp relief by Covid-19

The proposed system’s key features include:

• replacing annual tests for all pupils, with tests of a nationally representative sample of pupils conducted at key intervals
• new, broader types of assessment made freely available so that schools can use them to provide information for parents about their child’s achievements
• a new organisation to implement the system which would report directly to parliament
• radical reform of the inspection system so that the new research base generated from the national sample can be used to directly support school improvement
• informative reports allowing schools to compare themselves with others working in similar circumstances to aid system improvement.

Video of the report’s online launch event, featuring speakers and discussants including Margaret Greenwood MP, Pamela Sammons, Jo-Anne Baird and author Gemma Moss, is now online.

Report summary

In this report, the BERA Expert Panel on Assessment sets out a far-reaching yet pragmatic proposal to reform England’s high-stakes testing and accountability system in primary education and beyond, in ways that will benefit parents, pupils, teachers and policymakers.

We propose collecting richer information from a smaller number of pupils, for the express purpose of building knowledge for intelligent and democratic accountability. Replacing SATs (the sequence of national curriculum tests and checks that every child is expected to take in primary school between the ages of four and 11) with a national sample would eliminate the many drawbacks of the current use of high-stakes testing in the English education system while substantially improving the quality of pupils’ educational experiences. Our proposal would improve understanding of a broader range of factors that impact on pupils’ attainment, their wellbeing and system equity, while continuing to inform the general public and policymakers about the strengths and limitations of the education system as a whole.

Our proposed new system has seven key features.

1. A longitudinal sample of pupils, using the national pupil database (NPD) as a sampling frame, from whom additional data will be collected using a combination of assessment instruments and pupil/teacher/parent survey data. This combination of data will facilitate a richer exploration of educational processes and outcomes than the current system allows.
2. A new organisation to implement the system, acting independently of government and reporting directly to parliament. This body will take responsibility for stakeholder involvement, the design of instruments, the commissioning of the sample and reporting.
3. New assessment instruments, including but going beyond literacy and numeracy tests and their current narrow remits. These will be designed to focus on broader issues in pupils’ educational development. Once validated, assessment and survey instruments will be made public so that stakeholders can reuse them in light of the purposes for which they were designed. Sample items from a national bank of assessment instruments will enable schools to provide helpful information for parents about their child’s achievements.
4. Summary reports, based on the national sample and accessible to the public, will ensure that pupils, parents and schools benefit from the sample data to the fullest possible extent. The national data provided by this sample will help schools to evaluate their performance and reflect on relevant priorities for improvement in their context.
5. Different inspection processes, redesigned to encourage schools to use the research evidence from the national sample to inform actions in their own context. These will build on successful uses of professional and democratic processes of peer-to-peer review and self-evaluation to support school improvement in dialogue with other stakeholders and external partners.
6. Ongoing stakeholder involvement to help determine new topics of importance to be explored through the sample, and to foster the engagement of the research community. These may evolve over time to reflect new concerns (such as cyberbullying, mental health and the impact of Covid-19).
7. Once the new approach has been trialled and its benefits assessed, existing census-based annual testing (including SATs) of the whole school population in the primary and early years sectors will be phased out, saving pupil and teacher time, eliminating negative side-effects and reducing costs.

This approach would generate better system knowledge that, in conjunction with supportive and research-informed inspection, and could be drawn on by schools in the development of their own improvement plans. This combination of data would be better able to highlight issues outside of schools that may be crucial to addressing in-school performance, and link research to practice in mutually strengthening ways.

Recommendations

1. A new organisation to introduce & oversee a national sample

A new statutory body, independent of the executive and reporting directly to parliament, should be established to set the parameters and operational terms of reference of the new system. It would take responsibility for stakeholder involvement, the design of instruments, commissioning the sample and reporting. Its membership should be representative of different stakeholder groups: local authorities (LAs), multi-academy trusts, Ofsted and Ofqual, teachers’ unions and other professional associations and education experts drawn from learned societies, as well as parents and students. Such a broad membership would enable the building of a stronger and more stable school system, based on consultation with a full range of stakeholders and taking into account the needs of the system as a whole.

2. A national longitudinal sample designed to inform system improvement

The national sample would consist of pupils chosen in successive years for testing and subsequent follow-up at the equivalent of key stages 1, 2 and 3. The sampled cohort would be followed up at regular intervals so that, importantly, longitudinal relationships can be studied with respect to pupil characteristics – such as eligibility for free school meals or special educational needs and disabilities – that can change over time, allowing  fuller exploration of the factors that affect learning development. Furthermore, different samples of pupils could be selected for different purposes, including one-off studies that gather particular information of topical interest (such as the impacts of Covid-19). Pupils’ performance in public examinations would remain available for all candidates, meaning that contextual comparisons could still be made, and equity in outcomes studied, as required, with the national sample adding depth to those enquiries.

3. A new body to design assessment & survey instruments

To better describe learning development over time, a broader set of assessment instruments should replace the current narrow remit of SATs. This would form a national bank of high quality assessment instruments.

The bodies responsible for their design be independent of the government of the day, as the former Assessment and Performance Unit and Qualifications and Curriculum Authority once did. Testing would thereby revert to being uncoupled from political control.

The assessment instruments and the accompanying questionnaires used to collect relevant contextual information from the sample should be published, both in the spirit of transparency and to allow individual LAs, schools and other stakeholders to utilise items chosen from the national bank of assessment instruments.

4. A bank of assessment & survey instruments that can be shared

A well thought-through system of teacher assessment aligned with, but not identical to, sample items from a national bank of assessment instruments will enable schools to regularly provide helpful information for parents about their child’s achievements. Schools could use nationally standardised items to evaluate how their pupils are doing relative to other pupils in similar contexts. Importantly, such benchmarking assessments would be used to support school self-evaluation.

5. Inspection repurposed to inform school improvement

To make best use of the data throughout the education system, the roles and responsibilities of existing bodies responsible for assessment (the Standards and Testing Agency and Ofqual) and inspection (Ofsted) would need to be remodelled. Ofsted’s purposes and remit should change to focus more directly on school improvement, generating reports to parliament on how the system can be improved and its weaknesses addressed, rather than on the repeated grading of individual schools.